Skip to content

Latest commit

 

History

History
1186 lines (726 loc) · 42.2 KB

k8s.md

File metadata and controls

1186 lines (726 loc) · 42.2 KB

Documentation for scw k8s

This API allows you to manage Kubernetes Kapsule and Kosmos clusters.

Access Control List (ACL) management commands

Network Access Control Lists (ACLs) allow you to manage inbound network traffic by setting up ACL rules.

Network Access Control Lists (ACLs) allow you to manage inbound network traffic by setting up ACL rules.

Usage:

scw k8s acl

Kapsule cluster management commands

A cluster is a fully managed Kubernetes cluster It is composed of different pools, each pool containing the same kind of nodes.

Create a new Cluster

Create a new Kubernetes cluster in a Scaleway region.

Usage:

scw k8s cluster create [arg=value ...]

Args:

Name Description
project-id Project ID to use. If none is passed the default project ID will be used
type Type of the cluster (possible values are kapsule, multicloud, kapsule-dedicated-8, kapsule-dedicated-16)
name Required
Default: <generated>
Cluster name
description Cluster description
tags.{index} Tags associated with the cluster
version Required
Default: latest
Kubernetes version of the cluster
cni Required
Default: cilium
One of: unknown_cni, cilium, calico, weave, flannel, kilo, none
Container Network Interface (CNI) plugin running in the cluster
pools.{index}.name Name of the pool
pools.{index}.node-type Node type is the type of Scaleway Instance wanted for the pool. Nodes with insufficient memory are not eligible (DEV1-S, PLAY2-PICO, STARDUST). 'external' is a special node type used to provision instances from other cloud providers in a Kosmos Cluster
pools.{index}.placement-group-id Placement group ID in which all the nodes of the pool will be created
pools.{index}.autoscaling Defines whether the autoscaling feature is enabled for the pool
pools.{index}.size Size (number of nodes) of the pool
pools.{index}.min-size Defines the minimum size of the pool. Note that this field is only used when autoscaling is enabled on the pool
pools.{index}.max-size Defines the maximum size of the pool. Note that this field is only used when autoscaling is enabled on the pool
pools.{index}.container-runtime One of: unknown_runtime, docker, containerd, crio Customization of the container runtime is available for each pool. Note that docker has been deprecated since version 1.20 and will be removed by version 1.24
pools.{index}.autohealing Defines whether the autohealing feature is enabled for the pool
pools.{index}.tags.{index} Tags associated with the pool
pools.{index}.kubelet-args.{key} Kubelet arguments to be used by this pool. Note that this feature is experimental
pools.{index}.upgrade-policy.max-unavailable The maximum number of nodes that can be not ready at the same time
pools.{index}.upgrade-policy.max-surge The maximum number of nodes to be created during the upgrade
pools.{index}.zone Zone in which the pool's nodes will be spawned
pools.{index}.root-volume-type One of: default_volume_type, l_ssd, b_ssd Defines the system volume disk type. Two different types of volume (volume_type) are provided: l_ssd is a local block storage which means your system is stored locally on your node's hypervisor. b_ssd is a remote block storage which means your system is stored on a centralized and resilient cluster
pools.{index}.root-volume-size System volume disk size
pools.{index}.public-ip-disabled Defines if the public IP should be removed from Nodes. To use this feature, your Cluster must have an attached Private Network set up with a Public Gateway
autoscaler-config.scale-down-disabled Disable the cluster autoscaler
autoscaler-config.scale-down-delay-after-add How long after scale up that scale down evaluation resumes
autoscaler-config.estimator One of: unknown_estimator, binpacking Type of resource estimator to be used in scale up
autoscaler-config.expander One of: unknown_expander, random, most_pods, least_waste, priority, price Type of node group expander to be used in scale up
autoscaler-config.ignore-daemonsets-utilization Ignore DaemonSet pods when calculating resource utilization for scaling down
autoscaler-config.balance-similar-node-groups Detect similar node groups and balance the number of nodes between them
autoscaler-config.expendable-pods-priority-cutoff Pods with priority below cutoff will be expendable. They can be killed without any consideration during scale down and they won't cause scale up. Pods with null priority (PodPriority disabled) are non expendable
autoscaler-config.scale-down-unneeded-time How long a node should be unneeded before it is eligible to be scaled down
autoscaler-config.scale-down-utilization-threshold Node utilization level, defined as a sum of requested resources divided by capacity, below which a node can be considered for scale down
autoscaler-config.max-graceful-termination-sec Maximum number of seconds the cluster autoscaler waits for pod termination when trying to scale down a node
auto-upgrade.enable Defines whether auto upgrade is enabled for the cluster
auto-upgrade.maintenance-window.start-hour Start time of the two-hour maintenance window
auto-upgrade.maintenance-window.day One of: any, monday, tuesday, wednesday, thursday, friday, saturday, sunday Day of the week for the maintenance window
feature-gates.{index} List of feature gates to enable
admission-plugins.{index} List of admission plugins to enable
open-id-connect-config.issuer-url URL of the provider which allows the API server to discover public signing keys. Only URLs using the https:// scheme are accepted. This is typically the provider's discovery URL without a path, for example "https://accounts.google.com" or "https://login.salesforce.com"
open-id-connect-config.client-id A client ID that all tokens must be issued for
open-id-connect-config.username-claim JWT claim to use as the user name. The default is sub, which is expected to be the end user's unique identifier. Admins can choose other claims, such as email or name, depending on their provider. However, claims other than email will be prefixed with the issuer URL to prevent name collision
open-id-connect-config.username-prefix Prefix prepended to username claims to prevent name collision (such as system: users). For example, the value oidc: will create usernames like oidc:jane.doe. If this flag is not provided and username_claim is a value other than email, the prefix defaults to ( Issuer URL )# where ( Issuer URL ) is the value of issuer_url. The value - can be used to disable all prefixing
open-id-connect-config.groups-claim.{index} JWT claim to use as the user's group
open-id-connect-config.groups-prefix Prefix prepended to group claims to prevent name collision (such as system: groups). For example, the value oidc: will create group names like oidc:engineering and oidc:infra
open-id-connect-config.required-claim.{index} Multiple key=value pairs describing a required claim in the ID token. If set, the claims are verified to be present in the ID token with a matching value
apiserver-cert-sans.{index} Additional Subject Alternative Names for the Kubernetes API server certificate
private-network-id Private network ID for internal cluster communication (cannot be changed later). For Kapsule clusters, if none is provided, a private network will be created
organization-id Organization ID to use. If none is passed the default organization ID will be used
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Create a Kubernetes cluster named foo with cilium as CNI, in version 1.27.0 and with a pool named default composed of 3 DEV1-M

scw k8s cluster create name=foo version=1.27.0 pools.0.size=3 pools.0.node-type=DEV1-M pools.0.name=default

Create a Kubernetes cluster named bar, tagged, calico as CNI, in version 1.27.0 and with a tagged pool named default composed of 2 RENDER-S and autohealing and autoscaling enabled (between 1 and 10 nodes)

scw k8s cluster create name=bar version=1.27.0 tags.0=tag1 tags.1=tag2 cni=calico pools.0.size=2 pools.0.node-type=RENDER-S pools.0.min-size=1 pools.0.max-size=10 pools.0.autohealing=true pools.0.autoscaling=true pools.0.tags.0=pooltag1 pools.0.tags.1=pooltag2 pools.0.name=default

Delete a Cluster

Delete a specific Kubernetes cluster and all its associated pools and nodes. Note that this method will not delete any Load Balancer or Block Volume that are associated with the cluster.

Usage:

scw k8s cluster delete <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required ID of the cluster to delete
with-additional-resources Defines whether all volumes (including retain volume type), empty Private Networks and Load Balancers with a name starting with the cluster ID will also be deleted
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Delete a cluster

scw k8s cluster delete 11111111-1111-1111-111111111111

Delete a cluster with its Block volumes and Load Balancers

scw k8s cluster delete 11111111-1111-1111-111111111111 with-additional-resources=true

Get a Cluster

Retrieve information about a specific Kubernetes cluster.

Usage:

scw k8s cluster get <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required ID of the requested cluster
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Get a cluster information

scw k8s cluster get 11111111-1111-1111-111111111111

List Clusters

List all existing Kubernetes clusters in a specific region.

Usage:

scw k8s cluster list [arg=value ...]

Args:

Name Description
project-id Project ID on which to filter the returned clusters
order-by One of: created_at_asc, created_at_desc, updated_at_asc, updated_at_desc, name_asc, name_desc, status_asc, status_desc, version_asc, version_desc Sort order of returned clusters
name Name to filter on, only clusters containing this substring in their name will be returned
status One of: unknown, creating, ready, deleting, deleted, updating, locked, pool_required Status to filter on, only clusters with this status will be returned
type Type to filter on, only clusters with this type will be returned
private-network-id Private Network ID to filter on, only clusters within this Private Network will be returned
organization-id Organization ID on which to filter the returned clusters
region Default: fr-par
One of: fr-par, nl-ams, pl-waw, all
Region to target. If none is passed will use default region from the config

Examples:

List all clusters on your default region

scw k8s cluster list

List the ready clusters on your default region

scw k8s cluster list status=ready

List the clusters that match the given name on fr-par ('cluster1' will return 'cluster100' and 'cluster1' but not 'foo')

scw k8s cluster list region=fr-par name=cluster1

List available cluster types for a cluster

List the cluster types that a specific Kubernetes cluster is allowed to switch to.

Usage:

scw k8s cluster list-available-types <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID for which the available Kubernetes types will be listed
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

List all cluster types that a cluster can upgrade to

scw k8s cluster list-available-types 11111111-1111-1111-111111111111

List available versions for a Cluster

List the versions that a specific Kubernetes cluster is allowed to upgrade to. Results will include every patch version greater than the current patch, as well as one minor version ahead of the current version. Any upgrade skipping a minor version will not work.

Usage:

scw k8s cluster list-available-versions <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID for which the available Kubernetes versions will be listed
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

List all versions that a cluster can upgrade to

scw k8s cluster list-available-versions 11111111-1111-1111-111111111111

Reset the admin token of a Cluster

Reset the admin token for a specific Kubernetes cluster. This will revoke the old admin token (which will not be usable afterwards) and create a new one. Note that you will need to download kubeconfig again to keep interacting with the cluster.

Usage:

scw k8s cluster reset-admin-token <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID on which the admin token will be renewed
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Reset the admin token for a cluster

scw k8s cluster reset-admin-token 11111111-1111-1111-111111111111

Change the Cluster type

Change the type of a specific Kubernetes cluster. To see the possible values you can enter for the type field, list available cluster types.

Usage:

scw k8s cluster set-type <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required ID of the cluster to migrate from one type to another
type Required Type of the cluster. Note that some migrations are not possible (please refer to product documentation)
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Convert a kapsule cluster to a kapsule-dedicated-16 cluster

scw k8s cluster set-type 11111111-1111-1111-111111111111 type=kapsule-dedicated-16

Update a Cluster

Update information on a specific Kubernetes cluster. You can update details such as its name, description, tags and configuration. To upgrade a cluster, you will need to use the dedicated endpoint.

Usage:

scw k8s cluster update <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required ID of the cluster to update
name New external name for the cluster
description New description for the cluster
tags.{index} New tags associated with the cluster
autoscaler-config.scale-down-disabled Disable the cluster autoscaler
autoscaler-config.scale-down-delay-after-add How long after scale up that scale down evaluation resumes
autoscaler-config.estimator One of: unknown_estimator, binpacking Type of resource estimator to be used in scale up
autoscaler-config.expander One of: unknown_expander, random, most_pods, least_waste, priority, price Type of node group expander to be used in scale up
autoscaler-config.ignore-daemonsets-utilization Ignore DaemonSet pods when calculating resource utilization for scaling down
autoscaler-config.balance-similar-node-groups Detect similar node groups and balance the number of nodes between them
autoscaler-config.expendable-pods-priority-cutoff Pods with priority below cutoff will be expendable. They can be killed without any consideration during scale down and they won't cause scale up. Pods with null priority (PodPriority disabled) are non expendable
autoscaler-config.scale-down-unneeded-time How long a node should be unneeded before it is eligible to be scaled down
autoscaler-config.scale-down-utilization-threshold Node utilization level, defined as a sum of requested resources divided by capacity, below which a node can be considered for scale down
autoscaler-config.max-graceful-termination-sec Maximum number of seconds the cluster autoscaler waits for pod termination when trying to scale down a node
auto-upgrade.enable Defines whether auto upgrade is enabled for the cluster
auto-upgrade.maintenance-window.start-hour Start time of the two-hour maintenance window
auto-upgrade.maintenance-window.day One of: any, monday, tuesday, wednesday, thursday, friday, saturday, sunday Day of the week for the maintenance window
feature-gates.{index} List of feature gates to enable
admission-plugins.{index} List of admission plugins to enable
open-id-connect-config.issuer-url URL of the provider which allows the API server to discover public signing keys. Only URLs using the https:// scheme are accepted. This is typically the provider's discovery URL without a path, for example "https://accounts.google.com" or "https://login.salesforce.com"
open-id-connect-config.client-id A client ID that all tokens must be issued for
open-id-connect-config.username-claim JWT claim to use as the user name. The default is sub, which is expected to be the end user's unique identifier. Admins can choose other claims, such as email or name, depending on their provider. However, claims other than email will be prefixed with the issuer URL to prevent name collision
open-id-connect-config.username-prefix Prefix prepended to username claims to prevent name collision (such as system: users). For example, the value oidc: will create usernames like oidc:jane.doe. If this flag is not provided and username_claim is a value other than email, the prefix defaults to ( Issuer URL )# where ( Issuer URL ) is the value of issuer_url. The value - can be used to disable all prefixing
open-id-connect-config.groups-claim.{index} JWT claim to use as the user's group
open-id-connect-config.groups-prefix Prefix prepended to group claims to prevent name collision (such as system: groups). For example, the value oidc: will create group names like oidc:engineering and oidc:infra
open-id-connect-config.required-claim.{index} Multiple key=value pairs describing a required claim in the ID token. If set, the claims are verified to be present in the ID token with a matching value
apiserver-cert-sans.{index} Additional Subject Alternative Names for the Kubernetes API server certificate
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Add TTLAfterFinished and ServiceNodeExclusion as feature gates on a cluster

scw k8s cluster update 11111111-1111-1111-111111111111 feature-gates.0=TTLAfterFinished feature-gates.1=ServiceNodeExclusion

Upgrade a Cluster

Upgrade a specific Kubernetes cluster and possibly its associated pools to a specific and supported Kubernetes version.

Usage:

scw k8s cluster upgrade <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required ID of the cluster to upgrade
version Required New Kubernetes version of the cluster. Note that the version should either be a higher patch version of the same minor version or the direct minor version after the current one
upgrade-pools Defines whether pools will also be upgraded once the control plane is upgraded
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Upgrade a cluster version 1.27.0 of Kubernetes (pools are not included)

scw k8s cluster upgrade 11111111-1111-1111-111111111111 version=1.27.0

Upgrade a cluster to version 1.27.0 of Kubernetes (pools are included)

scw k8s cluster upgrade 11111111-1111-1111-111111111111 version=1.27.0 upgrade-pools=true

Wait for a cluster to reach a stable state

Wait for server to reach a stable state. This is similar to using --wait flag on other action commands, but without requiring a new action on the server.

Usage:

scw k8s cluster wait <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required ID of the cluster.
wait-for-pools Wait for pools to be ready.
region Default: fr-par Region to target. If none is passed will use default region from the config
timeout Default: 10m0s Timeout of the wait

Examples:

Wait for a cluster to reach a stable state

scw k8s cluster wait 11111111-1111-1111-1111-111111111111

Cluster type management commands

All cluster types available in a specified region A cluster type represents the different commercial types of clusters offered by Scaleway.

List cluster types

List available cluster types and their technical details.

Usage:

scw k8s cluster-type list [arg=value ...]

Args:

Name Description
region Default: fr-par
One of: fr-par, nl-ams, pl-waw, all
Region to target. If none is passed will use default region from the config

Manage your Kubernetes Kapsule cluster's kubeconfig files

Retrieve a kubeconfig

Retrieve the kubeconfig for a specified cluster.

Usage:

scw k8s kubeconfig get <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID from which to retrieve the kubeconfig
region Default: fr-par Region to target. If none is passed will use default region from the config

Examples:

Get the kubeconfig for a given cluster

scw k8s kubeconfig get 11111111-1111-1111-1111-111111111111

Install a kubeconfig

Retrieve the kubeconfig for a specified cluster and write it on disk. It will merge the new kubeconfig in the file pointed by the KUBECONFIG variable. If empty it will default to $HOME/.kube/config.

Usage:

scw k8s kubeconfig install <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID from which to retrieve the kubeconfig
keep-current-context Whether or not to keep the current kubeconfig context unmodified
region Default: fr-par Region to target. If none is passed will use default region from the config

Examples:

Install the kubeconfig for a given cluster and using the new context

scw k8s kubeconfig install 11111111-1111-1111-1111-111111111111

Uninstall a kubeconfig

Remove specified cluster from kubeconfig file specified by the KUBECONFIG env, if empty it will default to $HOME/.kube/config. If the current context points to this cluster, it will be set to an empty context.

Usage:

scw k8s kubeconfig uninstall <cluster-id ...> [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID from which to uninstall the kubeconfig

Examples:

Uninstall the kubeconfig for a given cluster

scw k8s kubeconfig uninstall 11111111-1111-1111-1111-111111111111

Kapsule node management commands

A node (short for worker node) is an abstraction for a Scaleway Instance A node is always part of a pool. Each of them has the Kubernetes software automatically installed and configured by Scaleway.

Delete a Node in a Cluster

Delete a specific Node. Note that when there is not enough space to reschedule all the pods (such as in a one-node cluster), disruption of your applications can be expected.

Usage:

scw k8s node delete <node-id ...> [arg=value ...]

Args:

Name Description
node-id Required ID of the node to replace
skip-drain Skip draining node from its workload (Note: this parameter is currently inactive)
replace Add a new node after the deletion of this node
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Delete a node

scw k8s node delete 11111111-1111-1111-111111111111

Delete a node without evicting workloads

scw k8s node delete 11111111-1111-1111-111111111111 skip-drain=true

Replace a node by a new one

scw k8s node delete 11111111-1111-1111-111111111111 replace=true

Get a Node in a Cluster

Retrieve details about a specific Kubernetes Node.

Usage:

scw k8s node get <node-id ...> [arg=value ...]

Args:

Name Description
node-id Required ID of the requested node
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Get a node

scw k8s node get 11111111-1111-1111-111111111111

List Nodes in a Cluster

List all the existing nodes for a specific Kubernetes cluster.

Usage:

scw k8s node list [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID from which the nodes will be listed from
pool-id Pool ID on which to filter the returned nodes
order-by One of: created_at_asc, created_at_desc Sort order of the returned nodes
name Name to filter on, only nodes containing this substring in their name will be returned
status One of: unknown, creating, not_ready, ready, deleting, deleted, locked, rebooting, creation_error, upgrading, starting, registering Status to filter on, only nodes with this status will be returned
region Default: fr-par
One of: fr-par, nl-ams, pl-waw, all
Region to target. If none is passed will use default region from the config

Examples:

List all the nodes in the cluster

scw k8s node list cluster-id=11111111-1111-1111-111111111111

List all the nodes in the cluster's 2222222222222-2222-222222222222 pool

scw k8s node list cluster-id=11111111-1111-1111-111111111111 pool-id=2222222222222-2222-222222222222

List all cluster nodes that are ready

scw k8s node list cluster-id=11111111-1111-1111-111111111111 status=ready

Reboot a Node in a Cluster

Reboot a specific Node. The node will first be cordoned (scheduling will be disabled on it). The existing pods on the node will then be drained and rescheduled onto another schedulable node. Note that when there is not enough space to reschedule all the pods (such as in a one-node cluster), disruption of your applications can be expected.

Usage:

scw k8s node reboot <node-id ...> [arg=value ...]

Args:

Name Description
node-id Required ID of the node to reboot
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Reboot a node

scw k8s node reboot 11111111-1111-1111-111111111111

Replace a Node in a Cluster

Replace a specific Node. The node will first be cordoned (scheduling will be disabled on it). The existing pods on the node will then be drained and rescheduled onto another schedulable node. Note that when there is not enough space to reschedule all the pods (such as in a one-node cluster), disruption of your applications can be expected.

Usage:

scw k8s node replace <node-id ...> [arg=value ...]

Args:

Name Description
node-id Required ID of the node to replace
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Replace a node

scw k8s node replace 11111111-1111-1111-111111111111

Wait for a node to reach a stable state

Wait for a node to reach a stable state. This is similar to using --wait flag on other action commands, but without requiring a new action on the node.

Usage:

scw k8s node wait <node-id ...> [arg=value ...]

Args:

Name Description
node-id Required ID of the node.
region Default: fr-par Region to target. If none is passed will use default region from the config
timeout Default: 10m0s Timeout of the wait

Examples:

Wait for a node to reach a stable state

scw k8s node wait 11111111-1111-1111-1111-111111111111

Kapsule pool management commands

A pool is a set of identical nodes A pool has a name, a size (its desired number of nodes), node number limits (min, max), and a Scaleway Instance type. Changing those limits increases/decreases the size of a pool. As a result and depending on its load, the pool will grow or shrink within those limits when autoscaling is enabled. A "default pool" is automatically created with every cluster via the console.

Create a new Pool in a Cluster

Create a new pool in a specific Kubernetes cluster.

Usage:

scw k8s pool create [arg=value ...]

Args:

Name Description
cluster-id Required Cluster ID to which the pool will be attached
name Required
Default: <generated>
Pool name
node-type Required
Default: DEV1-M
Node type is the type of Scaleway Instance wanted for the pool. Nodes with insufficient memory are not eligible (DEV1-S, PLAY2-PICO, STARDUST). 'external' is a special node type used to provision instances from other cloud providers in a Kosmos Cluster
placement-group-id Placement group ID in which all the nodes of the pool will be created
autoscaling Defines whether the autoscaling feature is enabled for the pool
size Required
Default: 1
Size (number of nodes) of the pool
min-size Defines the minimum size of the pool. Note that this field is only used when autoscaling is enabled on the pool
max-size Defines the maximum size of the pool. Note that this field is only used when autoscaling is enabled on the pool
container-runtime One of: unknown_runtime, docker, containerd, crio Customization of the container runtime is available for each pool. Note that docker has been deprecated since version 1.20 and will be removed by version 1.24
autohealing Defines whether the autohealing feature is enabled for the pool
tags.{index} Tags associated with the pool
kubelet-args.{key} Kubelet arguments to be used by this pool. Note that this feature is experimental
upgrade-policy.max-unavailable
upgrade-policy.max-surge
zone Zone in which the pool's nodes will be spawned
root-volume-type One of: default_volume_type, l_ssd, b_ssd Defines the system volume disk type. Two different types of volume (volume_type) are provided: l_ssd is a local block storage which means your system is stored locally on your node's hypervisor. b_ssd is a remote block storage which means your system is stored on a centralized and resilient cluster
root-volume-size System volume disk size
public-ip-disabled Defines if the public IP should be removed from Nodes. To use this feature, your Cluster must have an attached Private Network set up with a Public Gateway
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Create a pool named 'bar' with 2 DEV1-XL on a cluster

scw k8s pool create cluster-id=11111111-1111-1111-111111111111 name=bar node-type=DEV1-XL size=2

Create a pool named 'fish' with 5 GP1-L, autoscaling within 0 and 10 nodes and autohealing enabled, and containerd as the cluster container runtime

scw k8s pool create cluster-id=11111111-1111-1111-111111111111 name=fish node-type=GP1-L size=5 min-size=0 max-size=10 autoscaling=true autohealing=true container-runtime=containerd

Create a tagged pool named 'turtle' with 1 GP1-S which is using the already created placement group 2222222222222-2222-222222222222 for all the nodes in the pool on a cluster

scw k8s pool create cluster-id=11111111-1111-1111-111111111111 name=turtle node-type=GP1-S size=1 placement-group-id=2222222222222-2222-222222222222 tags.0=turtle tags.1=placement-group

Delete a Pool in a Cluster

Delete a specific pool from a cluster. Note that all the pool's nodes will also be deleted.

Usage:

scw k8s pool delete <pool-id ...> [arg=value ...]

Args:

Name Description
pool-id Required ID of the pool to delete
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Delete a specific pool

scw k8s pool delete 11111111-1111-1111-111111111111

Get a Pool in a Cluster

Retrieve details about a specific pool in a Kubernetes cluster.

Usage:

scw k8s pool get <pool-id ...> [arg=value ...]

Args:

Name Description
pool-id Required ID of the requested pool
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Get a given pool

scw k8s pool get 11111111-1111-1111-111111111111

List Pools in a Cluster

List all the existing pools for a specific Kubernetes cluster.

Usage:

scw k8s pool list [arg=value ...]

Args:

Name Description
cluster-id Required ID of the cluster whose pools will be listed
order-by One of: created_at_asc, created_at_desc, updated_at_asc, updated_at_desc, name_asc, name_desc, status_asc, status_desc, version_asc, version_desc Sort order of returned pools
name Name to filter on, only pools containing this substring in their name will be returned
status One of: unknown, ready, deleting, deleted, scaling, warning, locked, upgrading Status to filter on, only pools with this status will be returned
region Default: fr-par
One of: fr-par, nl-ams, pl-waw, all
Region to target. If none is passed will use default region from the config

Examples:

List all pools for a cluster

scw k8s pool list cluster-id=11111111-1111-1111-111111111111

List all scaling pools for a cluster

scw k8s pool list cluster-id=11111111-1111-1111-111111111111 status=scaling

List all pools for clusters containing 'foo' in their name

scw k8s pool list cluster-id=11111111-1111-1111-111111111111 name=foo

List all pools for a cluster and order them by ascending creation date

scw k8s pool list cluster-id=11111111-1111-1111-111111111111 order-by=created_at_asc

Update a Pool in a Cluster

Update the attributes of a specific pool, such as its desired size, autoscaling settings, and tags.

Usage:

scw k8s pool update <pool-id ...> [arg=value ...]

Args:

Name Description
pool-id Required ID of the pool to update
autoscaling New value for the pool autoscaling enablement
size New desired pool size
min-size New minimum size for the pool
max-size New maximum size for the pool
autohealing New value for the pool autohealing enablement
tags.{index} New tags associated with the pool
kubelet-args.{key} New Kubelet arguments to be used by this pool. Note that this feature is experimental
upgrade-policy.max-unavailable
upgrade-policy.max-surge
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Enable autoscaling on a given pool

scw k8s pool update 11111111-1111-1111-111111111111 autoscaling=true

Reduce the size and maximum size of a given pool to 4

scw k8s pool update 11111111-1111-1111-111111111111 size=4 max-size=4

Modify the tags of a given pool

scw k8s pool update 11111111-1111-1111-111111111111 tags.0=my tags.1=new tags.2=pool

Upgrade a Pool in a Cluster

Upgrade the Kubernetes version of a specific pool. Note that it only works if the targeted version matches the cluster's version.

Usage:

scw k8s pool upgrade <pool-id ...> [arg=value ...]

Args:

Name Description
pool-id Required ID of the pool to upgrade
version Required New Kubernetes version for the pool
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Upgrade a specific pool to the Kubernetes version 1.27.0

scw k8s pool upgrade 11111111-1111-1111-111111111111 version=1.27.0

Wait for a pool to reach a stable state

Wait for a pool to reach a stable state. This is similar to using --wait flag on other action commands, but without requiring a new action on the node.

Usage:

scw k8s pool wait <pool-id ...> [arg=value ...]

Args:

Name Description
pool-id Required ID of the pool.
region Default: fr-par Region to target. If none is passed will use default region from the config
timeout Default: 10m0s Timeout of the wait

Examples:

Wait for a pool to reach a stable state

scw k8s pool wait 11111111-1111-1111-1111-111111111111

Available Kubernetes versions commands

A version is a vanilla Kubernetes version like x.y.z It comprises a major version x, a minor version y, and a patch version z. At the minimum, Kapsule (Scaleway's managed Kubernetes), will support the last patch version for the past three minor releases. Also, each version has a different set of CNIs, eventually container runtimes, feature gates, and admission plugins available. See our Version Support Policy.

Get a Version

Retrieve a specific Kubernetes version and its details.

Usage:

scw k8s version get <version-name ...> [arg=value ...]

Args:

Name Description
version-name Required Requested version name
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

Get the Kubernetes version 1.27.0

scw k8s version get 1.27.0

List all available Versions

List all available versions for the creation of a new Kubernetes cluster.

Usage:

scw k8s version list [arg=value ...]

Args:

Name Description
region Default: fr-par
One of: fr-par, nl-ams, pl-waw
Region to target. If none is passed will use default region from the config

Examples:

List all available Kubernetes version in Kapsule

scw k8s version list